alamb commented on issue #12819:
URL: https://github.com/apache/datafusion/issues/12819#issuecomment-2444190036

   > It would be great to figure out a way to pass this nullability info (or 
return) the invoke at execution time. Do you think any of the suggestion in the 
initial post would be a good idea?
   
   I worry about making too many different combinations of invoke / 
invoke_no_args / invoke_nullable, etc
   
   > 
   > It seems like the nullability wasn't fully considered for function since 
the `is_nullable` and `return_type_from_exprs` where added later, would be 
great have this info throughout the trait.
   
   I agree nullability was not fully considered and having a proper nullability 
story would be good. 
   
   I think one thing that happened is that much of DataFusion's code assumes 
the input can be nullable and then switches to special case no-null code at 
runtime based on 
[`Array::null_count`](https://docs.rs/arrow/latest/arrow/array/trait.Array.html#method.null_count)
 rather than relying on the schema 
   
   That made ensuring that the schema's nullability aligned wasn't as important 
and so there are a bunch of places which are fast and loose with it


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org

Reply via email to